The End of Ordinary Scarcity

Star Trek hid production inside a wall.

A person asked for tea, and the tea appeared. No mines, farms, factories, roads, ships, warehouses, engineers, cleaners, repair crews, planning systems, energy grid, property law, or political fight over who controlled the machine.

That made the replicator feel like magic. A humanoid robot that can do any physical job is like a real-life replicator, but larger, uglier, and forced to run the hidden supply chain itself.

A machine that can do general physical work can mine ore, refine material, run factories, lay cables, pour concrete, build houses, build solar farms, maintain turbines, fabricate chips, operate data centres, repair other robots, and build the next generation of robots. Once that loop closes, human labour stops being the scarce input that organises society. The consequences of this are both easy to overstate and easy to understate. The world does not end. There will be life after work, but not as we know it.

The Wage Bargain Breaks

People earn money by selling their labour. Firms earn revenue because customers have income and spend it on goods and services. Customers have income because firms still need people to operate. This cycle holds true as long as labour is required, but robots are breaking that contract by making production less and less dependent on human labour.

The early effects look ordinary: higher productivity, and some roles shrinking while others grow. We have lived through versions of this before and absorbed them, if with some friction. If machines can perform most physical and cognitive work, wages will no longer be the primary means by which ordinary people obtain purchasing power. The problem will move from producing enough goods to deciding who gets the remaining scarce and desirable assets.

Society can patch the gap for a while with subsidised jobs, automation taxes, nationalised assets, or new public funds, and some of those patches may be necessary and good. None of them preserve the old cycle long-term.

A robot tax is the first answer people reach for. Bill Gates argued for one in a 2017 Quartz interview, suggesting that companies should be taxed when robots take over work that would otherwise have generated wages and payroll taxes.¹ Using that tax, the state gives people money. People spend it on goods produced by automated firms. The money flows back to whoever owns the robots, and the state then has to recover enough of it through taxes to keep the cycle running. That buys time but doesn’t solve the underlying problem. If a small class owns the automated production base, the cycle becomes a subsidy to that class. The public receives income; the owners receive revenue. The owners then use that revenue to buy influence, lower their taxes, shape regulation, and convert industrial ownership into political power. Each cycle allows them to extract more from the system and accumulate more of the few remaining scarce assets.

For this reason, if robots produce ordinary goods at scale, money loses much of its role as a production signal. Food, clothing, furniture, tools, medicine, transport, entertainment, maintenance, education, and many kinds of care can become cheap enough that charging for them starts to look like charging for air. Society can make everyone materially comfortable, but if it fails to address ownership and position, it will make objects cheap and keep dignity expensive. Power will move into land, access, law, and status.

The Last Scarcity

Some scarcities yield to more production. Energy looks scarce until machines can build nuclear plants, solar farms, grid storage, substations, transmission lines, and the factories that make the parts. Compute looks scarce until machines can build chip fabs, cooling systems, data centres, and the industrial base that supports them. Housing looks scarce until machines can produce materials, lay roads, fit plumbing and wiring, and build at speeds no human trade workforce can match. Food yields to automated farming, processing, transport, and preparation. Much of healthcare changes as diagnosis, drug discovery, surgery, monitoring, and elder care become heavily automated. Education changes when every person has a patient, expert tutor who teaches in the style they understand.

None of this is automatic. Raw materials, ecological limits, permitting, war, and institutional failure still matter. Politics can block abundance, regulation can slow it, incumbents can sabotage it, and bad institutions can turn cheap production back into artificial scarcity. These remain hard systems. They are also industrial problems, and machines that build more machines can address them.

Position is different. You can build more houses, but not more of the best locations. You can build more cars, but you cannot make a rare car common; once it is common, its rarity is gone. You can build more yachts and still not give everyone the best berth in the harbour. You can make commercial flights so good that private jets lose their value, and still not give everyone the status of skipping the queue. You can build artificial beaches, lakes, islands, and resorts, and none of them reproduce that coastline, that view, that neighbourhood, that distance from crime, that school catchment, or that climate. Every human body still has to be somewhere, and somewhere is finite. Someone gets the coast, someone gets the centre, someone gets the quiet road, and someone gets pushed out.

The corporate abundance story stays incomplete for that reason. Tesla’s 2025 Master Plan Part IV builds its pitch around “sustainable abundance” and puts AI, robotics, and physical-world automation at the centre.² It says the quiet part plainly: the ambition is no longer a better car or a better warehouse; it is an automated civilisation. That ambition is worth taking seriously, and so is the question the plan skips: who owns the civilisation once it runs itself?

The replicator did not give Star Trek a world without hierarchy. The characters still competed for command, reputation, ships, missions, quarters, relationships, and access to events. They did not fight over dinner, because production was a solved problem; instead, they fought over rank, recognition, loyalty, ideals, and power. That is closer to the future this technology points towards, because a robot economy would not stop humans from wanting things; it would only make ordinary things easy to get and push the conflict towards things that cannot be mass-produced.

The Robot Apocalypse

There is a weak version of this argument in which the robots arrive, and everything collapses. Jobs vanish, money loses its value, cities burn, and people retreat to bunkers. Before the COVID-19 pandemic, pandemic risk lived partly in serious public health work and partly in prepper imagination, and then a real pandemic arrived. It was bad. Lockdown was miserable, people died, institutions failed in visible ways, and governments made decisions people will argue over for decades. The world still did not end. Vaccines were authorised by national regulators and validated by the WHO within a year.³ Governments improvised income support, companies changed working patterns within weeks, and supply chains bent and broke while keeping moving. The system took a serious shock without turning into a bunker fantasy.

That cuts both ways. Human systems are more resilient than alarmists think and less stable than optimists assume. COVID did not end the world, but it did change the assumptions underlying ordinary life: remote work became normal where possible, supply chains stopped being invisible, and public health, borders, offices, and state power all became live political questions. AI robotics is a larger shock because it does not interrupt work for a season; it reduces the need for work across the whole system. The transition will be messy rather than cinematic. Expect robot taxes, income trials, public ownership proposals, sovereign AI funds, licensing regimes, protectionist laws, court cases, black markets, union campaigns, manufactured scarcity, real scarcity, corporate lobbying, moral panic, and obvious scams, often all at the same time. The wage economy will not vanish on a single Tuesday. Systems rarely collapse on a schedule. They mutate, absorb the pressure, and preserve entrenched advantages.

Work Does More Than Produce

Treating this as a jobs problem makes it too small. Work does several jobs at once: it produces goods, it distributes income, it gives people structure and status, it creates dependency, it hands the state a tax base, and it gives politics a moral language for deciding who deserves comfort and who deserves punishment. When production no longer needs most people, each of those other functions remains and still needs an answer. People still need income unless ordinary goods become public infrastructure. They still need status unless society finds non-economic ways to recognise contribution. They still need independence from families, partners, employers, towns, and local power, and a way to build a life that is not handed to them as charity by whoever owns the machines.

The robot future exposes how much moral weight we force wages to carry. A pay cheque did more than buy things. It was permission to exist without asking, proof of usefulness, bargaining power, and a way out of a bad situation. A world after wage labour has to replace all of that, or it produces a comfortable dependency that looks humane from a distance and rots up close.

Removing Excuses

The future should be materially better. People should not spend their lives doing dangerous, boring, repetitive work just to survive, and poverty should become harder to justify as machines make production cheap. Food, shelter, healthcare, education, transport, energy, and basic comfort should move closer to infrastructure than to luxury, and that would be a civilisational achievement. It would also remove an excuse. For centuries, we have rationed life through money and called the result natural, insisting there was not enough food, housing, care, time, or productive capacity to go round. Some of that scarcity was real, some was political, and some was maintained because it suited those in power.

Robots will not settle that argument on their own, but they may make it impossible to hide. Once production is cheap, every hierarchy has to defend itself without leaning so hard on necessity. The owner of the machine cannot say everyone is poor because there is not enough. The landlord cannot pretend that all exclusions are a production problem. The state cannot treat wages as the only legitimate means of conferring independence.

Replicators do not end politics. They remove one of the oldest alibis politics has ever had. The machines may make almost everything, and then we will find out what humans still refuse to share.